NU021172@NDSUVM1.BITNET (Marty Hoag) (12/10/87)
Steve Druck posted some comments on the IBMTCP-L list which I think
are relevant here and which probably are shared by many customers,
especially in Higher Education. Steve has granted me permission to
post the note here for further comments. Since Steve is not a member
of the list you might want to include him in any mail or send directly
to him.
I apologize for sending duplicates to those of you on both IBM7171 and
IBM-NETS lists. But this is one of those "cross-list" topics...
Marty
------------------------- Original Note from IBMTCP-L ----------------------
Date: Wed, 09 Dec 87 08:53:37 +0200
From: Steve Druck <VUSTEVE@WEIZMANN>
Subject: TCP/IP: IBM's Strategic Directions and Ours
I have been recently playing with IBM's GDDM PC Link. This product allows
GDDM graphics on PCs having ordinary PC graphics cards like the EGA,
or the new VGA in the PS/2. It also supports graphic printers or
plotters attached to the PC.
I also plan to test soon IBM's Enhanced Connectivity Facility, which
promises to provide mainframe support for PC virtual disks, virtual
printers, virtual files (virtually anything?).
The problem is that both of these products require either the use of
3270 emulation cards and IBM's Emulation Program in the PC
or the connection of several PCs through a PC running IBM's PC Workstation
Program. Our strategic plans are, however:
1) high speed connection of new PCs and workstations via TCP/IP and
the campus Ethernet to IBM mainframes either through a DACU
(later 8232) or, for TELNET, through a Bridge CS/100 (in
"milking-machine" mode) and a 7171.
2) lower speed (up to 19.2Kb) connections to IBM mainframes through
a digital switch to a 7171.
According to IBM, in neither fashion can we support PC Link and ECF. The
7171 terminal emulation will not support them. tn3270 and FAL won't
support them. I even checked if the asynchronous adapter feature in the 3174
will do so (answer: no).
I believe that our campus is not atypical. From what I see being discussed
in this list, it certainly isn't. Is IBM marching to a different drummer?
If these two products are an indication of IBM's strategic direction, will
IBM be marching alone?
While Token Ring-based networks may be technically superior to Ethernets,
I don't see too many around and a heck of a lot of TCP/IP-based
Ethernets have gone in over the last several years. At the very least,
IBM appears not to be providing a couple of items of strategic software
to a significant market segment. At worst, it may lose this market
segment because of its refusal to incorporate it in its overall strategy.